test indicator misadventure
Dave Martindale
curiosity, I removed the side plate from it to look at the internal
mechanism. This was a mistake, as it turns out that the side plate is an
integral part of the mechanism, carrying two jewel bearings and a
spring. Now, I've taken the side of a cheap test indicator before, and
on that one it was just a cover - all the internal mechanism was
undisturbed by cover removal. But not so with the Mitutoyo.
With the mechanism open, all the pieces still remained in place so I was
still in pretty good shape. The sector gear and crown gear were now
sitting there with only one end of their shafts in a bearing, but they
continued to sit in the right place. All I had to do was figure out how
to put it back together. This is a bit complicated because there's a
thin wire spring attached to the inside of the cover that apparently
provides much of the force that presses the contact tip onto the work
being measured. It seems to press on a pin which is part of the long
sector gear. At least, I think it must do that, because if it doesn't
then it doesn't touch anything at all inside the indicator.
Unfortunately, while trying to figure out what the spring does and how
to put it back into place while closing the mechanism up, I dropped the
indicator. (I was holding it in my hands, when I really should have
clamped it to a base and arm - mistake #2). It fell only a few inches
onto a padded mat on my desk - no dents or lost parts. But the sector
gear fell out of its place, which released the crown gear, which allowed
the spring-loaded dial to unwind completely.
Now, some terminology so everyone knows what I'm talking about. Some
terms are borrowed from a Mitutoyo catalog; I don't know if they are the
standard ones. There are 4 major moving assemblies in my dial indicator:
First, the contact point rotates on a pair of bearings at the end of the
indicator. There is also a lever that extends into the inside of the
indicator, moving in step with the contact point on its arm. There is a
friction coupling between the contact point and the inside lever portion
that will allow the contact point angle to be adjusted, but during
measuring the two of these move as a single unit.
Next, just inside the indicator is a "sector gear". This is a long arm
with its pivot point near one end. The lever arm from the contact point
presses against a pair of pins on the sector gear, causing the sector
gear to move whenever the contact point moves. This is also the
automatic reversing mechanism - no matter which direction the contact
point moves away from "neutral", the sector gear always rotates in the
same direction, due to the arrangement of the pins and the lever pushing
on them. (Really a clever mechanism). Also, there is a spring inside
the cover that tries to return the sector gear to its neutral position.
The sector gear *is* actually a gear, because there are a dozen or so gear
teeth cut in the far end of the long arm of the sector gear. If it was
a full circular gear, it would be something like 2 inches in diameter.
But it never turns more than a few degrees, so only a tiny section of
the outer portion of the gear is present, and the overall shape is that
of a long lever.
The teeth on the sector gear mesh with teeth on the small-diameter end of
the crown gear, so motion of the sector gear rotates the crown gear.
The large end of the crown gear meshes with a small pinion gear on the
dial shaft, and that moves the hand in the dial.
Now, the dial mechanism is preloaded with some kind of spring to take
up any slack in the gearing. This also provides some of the contact
force at the external point, but apparently not a lot of it. When I
dumped the sector gear out, I lost all this preload tension, and I don't
know how much of it there was.
So, my question: does anyone know how much rotation is necessary to
properly preload the dial in a dial test indicator? What I need to do is
rotate the crown gear the right number of turns to load this spring
before reinstalling the sector gear.
I did some experimenting, and found that I could rotate the pointer at
least 6 or 7 turns without breaking anything, while the actual operating
range of the pointer is 1.5 turns. About 7 turns of the pointer is one
turn of the crown gear. I found that cranking in 2 dial pointer turns
of preload just didn't seem enough tension - it wouldn't reliably
overcome friction in the gear train without the "extra" spring acting on
the sector gear. 3 turns of preload *does* seem to be enough to
overcome friction on its own, so that's what I used when I finally
reassembled it - the smallest amount of preload that seems like it
would work reliably. This ends up being about a half-turn of the crown
gear to do the preloading.
The reassembled indicator seems to work fine and has "reasonable"
contact pressure, so I think it's OK. But I'd really like to know
either the "official" way to set the preload, or get some idea of the
right way to guess it, finding the right balance between too little
(and possible errors due to allowing backlash to creep in) and too much
(increasing friction and wear due to higher-than-intended spring
forces).
Dave
Dave Baker
>From: da...@cs.ubc.ca (Dave Martindale)
>Date: 10/12/00 05:18 GMT Standard Time
>Message-id: <90v3nv$and$1...@lily.cs.ubc.ca>
>
>I recently acquired a Mitutoyo "quick set" dial test indicator. Out of
>curiosity, I removed the side plate from it to look at the internal
>mechanism. This was a mistake
<snip>
If you do a search in the Yellow Pages for metrology specialists you should be
able to find someone who rebuilds dial guages for a living and they should be
able to tell you how many turns of the spring are needed on this model.
http://members.aol.com/pumaracing/index.htm
"How do you feel?" asked Ford. "A bit like a military academy" replied Arthur,
"bits of me keep on passing out."
Spehro Pefhany
> contact pressure, so I think it's OK. But I'd really like to know
> either the "official" way to set the preload, or get some idea of the
> right way to guess it...
If you could compare it with a known-good similar one..
Best regards,
--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Spehro Pefhany --"it's the network..." "The Journey is the reward"
sp...@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com
Contributions invited->The AVR-gcc FAQ is at: http://www.BlueCollarLinux.com
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
mull...@advinc.com
da...@cs.ubc.ca (Dave Martindale) wrote:
> I recently acquired a Mitutoyo "quick set" dial test indicator...
[snip the most amazingly lucid text description of the insides
of dial indicator I've ever read, you sir must be a technical
writer or some such!]
When I took apart my starrett dial caliper, I did what Sphero
said - I finally resorted to opening up another similar one
to count the pre-load off.
In this case it was about seven teeth. Here this is an anti-backlash
idler gear that is between the rack pinion and the gear on the
dial pointer. The preload was provide by a tiny spring in a pocket
between two halves of the gear.
I can only suggest:
1) call Mitutoyo and ask the specific question of how much preload?
2) take apart another one, mark the parts before disassembly, and
count it that way.
3) If the pre-load is too little, the indidicator will be uncertain
as to how it should rest - the pointer will not decide where to
sit if you repeat the same measurement, say with the indicator
clamped on the toolpost and the crossfeed handle brought to the
same number each time. It should hit the same spot on the dial
each time, and if it is not, I would suspect the pre-load is
too little.
4) My experience shows if you have too much pre-load, the mechanism
will be noticeably stiff and sticky.
5) A good binocular microscope makes this stuff a lot easier!
If you have cleaned any of the parts then they might require oil
at the pivot points. Trick here is to not use too much. I use
automatic transmission fluid, applied with a 5 mil wire taped
to a stick, and the end of teh wire bend into a tiny hook to hold
a small amount of oil.
Jim
Sent via Deja.com http://www.deja.com/
Before you buy.
Ted Edwards
> Now, the dial mechanism is preloaded with some kind of spring to take
> up any slack in the gearing.
Exactly. Anti-backlash.
> This also provides some of the contact
> force at the external point, but apparently not a lot of it.
Non-zero but negligable on any dial indicator I ever had apart. Its
purpose is as you stated.
> I did some experimenting, and found that I could rotate the pointer at
> least 6 or 7 turns without breaking anything, while the actual operating
> range of the pointer is 1.5 turns. About 7 turns of the pointer is one
> turn of the crown gear. I found that cranking in 2 dial pointer turns
> of preload just didn't seem enough tension - it wouldn't reliably
> overcome friction in the gear train without the "extra" spring acting on
> the sector gear. 3 turns of preload *does* seem to be enough to
> overcome friction on its own, so that's what I used when I finally
> reassembled it - the smallest amount of preload that seems like it
> would work reliably. This ends up being about a half-turn of the crown
> gear to do the preloading.
I probably would have gone for four if the machanism is clean and
*properly* lubricated. I can't see that that would hurt. I haven't
done a Mit but I have done a couple of Federal and a few Starretts.
Ted
Dave Martindale
>> But I'd really like to know
>> either the "official" way to set the preload, or get some idea of the
>> right way to guess it...
>If you could compare it with a known-good similar one..
Trouble is, I don't know where I can borrow another one that someone would
allow me to take apart. And I don't fancy *buying* another one just to
take it apart and count preload turns, since I don't need a second one.
That's why I asked here, in case someone already knows the answer.
Dave
Dave Martindale
> [snip the most amazingly lucid text description of the insides
> of dial indicator I've ever read, you sir must be a technical
> writer or some such!]
Well, thank you. I'm not a technical writer, but I do occasionally
teach stuff to classes.
>When I took apart my starrett dial caliper, I did what Sphero
>said - I finally resorted to opening up another similar one
>to count the pre-load off.
Trouble is, I don't have another one, don't know where I can borrow one,
and don't want to buy a matching one. This is the single most expensive
piece of mechanical measuring equipment I own at the moment.
>In this case it was about seven teeth. Here this is an anti-backlash
>idler gear that is between the rack pinion and the gear on the
>dial pointer. The preload was provide by a tiny spring in a pocket
>between two halves of the gear.
In this case, was the preload applied between two different gears on
the same shaft? Or was this a single gear made up of two side-by-side
gears of the same diameter and tooth count?
>1) call Mitutoyo and ask the specific question of how much preload?
That's an interesting thought, if they would tell me (rather than just
"send it to us for repair").
>2) take apart another one, mark the parts before disassembly, and
> count it that way.
Doesn't seem practical - see above.
>3) If the pre-load is too little, the indidicator will be uncertain
> as to how it should rest - the pointer will not decide where to
> sit if you repeat the same measurement
>4) My experience shows if you have too much pre-load, the mechanism
> will be noticeably stiff and sticky.
The reassembled indicator behaves very well. Measurements seem
repeatable. Yet the needle shows no sign of stickiness - it will
easily show ripples in the surface of a piece of ribbon metal pulled
under the contact point, though the ripples are only 0.0001 high -
this is 1/5 of the smallest division on the dial. (The metal ribbon
is actually from a store anti-theft tag).
>5) A good binocular microscope makes this stuff a lot easier!
I don't have that, but I *did* use a Luxo illuminated magnifier,
sometimes a 5X loupe, and tweezers - basically the same tools I use
working with surface-mount electronic components.
>If you have cleaned any of the parts then they might require oil >at
the pivot points. Trick here is to not use too much. I use
>automatic transmission fluid, applied with a 5 mil wire taped
>to a stick, and the end of teh wire bend into a tiny hook to hold
>a small amount of oil.
I did add a bit of oil to the pivots - Starrett Tool & Instrument Oil
(seemed like a good choice) applied with a toothpick (not as good
control of amount as your method). I also applied some where the main
spring applies force to the sector gear, since that seemed like it
might involve some sliding contact. I didn't put any on the gear teeth
themselves, since there was no sign of oil on them initially.
---------------------------------------------------------------------
Hmm. I just thought of another way to (roughly) check whether I got
the preload right. Although the dial preload doesn't produce most of
the force on the measuring tip, it is responsible for some of it.
If the preload is wrong by a whole turn or more, the contact force may
be off by a measurable amount. If I could somehow borrow an identical
indicator, I could place their tips together and start pushing one with
the other. If both of the dials move at about the same time, the
measuring force is the same for both. If one of them moves well before
the other moves off its rest position, the one that moves first has the
lower contact force, and probably less dial preload. If one moves its
full 1.5 turns and hits the stop before the other one even starts, then
there is probably at least one and likely more than 1 turn difference
in the amount of preload. And this doesn't require taking apart the
second indicator at all.
Dave
mull...@advinc.com
da...@cs.ubc.ca (Dave Martindale) wrote:
> In [your] case, was the preload applied between two different gears on
> the same shaft? Or was this a single gear made up of two side-by-side
> gears of the same diameter and tooth count?
The starrett gear in question was the latter type, really a single
gear that is split in half, with a spring between the two halves.
> The reassembled indicator behaves very well. Measurements seem
> repeatable. Yet the needle shows no sign of stickiness - it will
> easily show ripples in the surface of a piece of ribbon metal pulled
> under the contact point, though the ripples are only 0.0001 high -
> this is 1/5 of the smallest division on the dial. (The metal ribbon
> is actually from a store anti-theft tag).
This is probably the best test for your assembly technique. If
you set it somewhere in the middle of the range, it should be fine.
Spehro Pefhany
> Trouble is, I don't know where I can borrow another one that someone would
> allow me to take apart. And I don't fancy *buying* another one just to
> take it apart and count preload turns, since I don't need a second one.
Right. I was just suggesting you compare the pressure to see if you got it
right already.
Ted Edwards
> In this case, was the preload applied between two different gears on
> the same shaft? Or was this a single gear made up of two side-by-side
> gears of the same diameter and tooth count?
My Federals have a hairspring on one of the gears. You can just make it
out on the gear in the lower right in photo dgage2.jpg of my writeup at
http://www.metalworking.com/dropbox/_1999_retired_files/DGAGE.TXT I set
the preload one turn more than necessary to prevent backlash after
cleaning and lubrication.
> The reassembled indicator behaves very well. Measurements seem
> repeatable. Yet the needle shows no sign of stickiness - it will
> easily show ripples in the surface of a piece of ribbon metal pulled
> under the contact point, though the ripples are only 0.0001 high -
> this is 1/5 of the smallest division on the dial. (The metal ribbon
> is actually from a store anti-theft tag).
Can't ask much better than that.
> I did add a bit of oil to the pivots - Starrett Tool & Instrument Oil
Good choice. 3-in-1 is a bad choice. It's called 3-in-1 because it's a
blend of three oils, light, medium and heavy. With time, the light
evaporates and leaves something much too heavy. I would advise against
using any automotive oils.
> (seemed like a good choice) applied with a toothpick (not as good
> control of amount as your method).
A fine sewing needle stuck in a piece of dowel with the point out makes
an excellent oiling tool. Don't put a hook in the end. This holds too
much oil especially for jewel bearings. It is truly amazing how
*little* oil is the right amount!
> be off by a measurable amount. If I could somehow borrow an identical
> indicator, I could place their tips together and start pushing one with
> the other. If both of the dials move at about the same time, the
> measuring force is the same for both.
Maybe but I wouldn't be surprised if the tollerance in the plunger
springs was more force than the anti-backlash spring supplies. After
all, the anti-backlash spring just has to overcome the friction in some
pretty good bearings.
I experimented with my Federals and found that there was several turns
tollerance between insufficient tension to overcome stickiness and
enough to cause the spring to shift out of plane.
Ted
mull...@advinc.com
Te...@bc.sympatico.ca wrote:
> I would advise against
> using any automotive oils.
ATF was originally formulated (supposedly) for replacing sperm
whale oil, which was a favorite for watch folks. I've used
it for years in indicators and watches, and all those that I've
re-built have given continued good service. It's formulated to
be non-gumming and has about the right viscosity.
But of course the starrett stuff is probably a good choice as well.
> A fine sewing needle stuck in a piece of dowel with the point out
makes an excellent oiling tool. Don't put a hook in the end. This
holds too much oil especially for jewel bearings.
The hook that I use is pretty small - and its more of an "L"
shape, really. The wire is about 5 mills, and the length that
I bend up with tweezers to get the hook is about the width of
the wire.
I find with straight oil tools, the oil forms a bead about a
couple of diameter up the tool, from surface tension. Then it
becomes an 'all or nothing' proposition. Made properly, the
kink, or L, or hook (whatever one calls it) delivers about the
right amount for a single pivot.
Dave Martindale
>My Federals have a hairspring on one of the gears. You can just make it
>out on the gear in the lower right in photo dgage2.jpg of my writeup at
>http://www.metalworking.com/dropbox/_1999_retired_files/DGAGE.TXT I set
>the preload one turn more than necessary to prevent backlash after
>cleaning and lubrication.
Ok. This is a test indicator, not a dial indicator, and I never did
attempt to take apart the dial mechanism, so I never saw the spring
inside the dial. I'd guess it is a hairspring, though.
What I found is that 2 turns of preload left the mechanism "sticky",
without enough force to return to zero reliably, if the main spring
that bears on the sector gear was disconnected and the dial anti-
backlash spring was providing all of the force to move everything. 3
turns of preload on the dial spring was enough to move the whole
mechanism smoothly. So that's where I left it.
This is actually conservative, because in normal operation the dial
spring only has to provide enough force to move the dial and the crown
gear, while the stronger main spring provides the force to move the
sector gear and the contact point mechanism. The 3 turns of preload
was enough to move all 4 moving parts without help from the main spring.
>3-in-1 is a bad choice. It's called 3-in-1 because it's a
>blend of three oils, light, medium and heavy. With time, the light
>evaporates and leaves something much too heavy.
Sounds like, no matter what the purpose, 3-in-1 is not the best oil for it.
>I would advise against using any automotive oils.
Particularly detergent oil - don't want something that actively picks up
water.
>A fine sewing needle stuck in a piece of dowel with the point out makes
>an excellent oiling tool. Don't put a hook in the end. This holds too
>much oil especially for jewel bearings. It is truly amazing how
>*little* oil is the right amount!
How much is the right amount? I put enough on the portion of the shaft
that goes into the bearing that I could see it sitting there with a loupe.
Not a huge blob, but more than would be needed to fill the space between
the steel and the jewel. What is the effect of using too much oil?
>I experimented with my Federals and found that there was several turns
>tollerance between insufficient tension to overcome stickiness and
>enough to cause the spring to shift out of plane.
In this case, I think I will avoid taking apart the dial mechanism, so
I can't see the spring.
Dave
Ted Edwards
> Sounds like, no matter what the purpose, 3-in-1 is not the best oil for it.
I don't have any at all either in the shop or house. We use instrument
oil on my wife's sewing machine.
> Particularly detergent oil - don't want something that actively picks up
> water.
For sure!
> How much is the right amount? I put enough on the portion of the shaft
> that goes into the bearing that I could see it sitting there with a loupe.
> Not a huge blob, but more than would be needed to fill the space between
> the steel and the jewel. What is the effect of using too much oil?
Sounds good. Too much seems to attract and hold dirt. Also spreads
itself around and fouls things that shouldn't be oiled. In the old
balance wheel watches, oil on the wheel would change its moment of
inertia. That's not good for accuracy. :-) The finest jewel bearings
consist of two parts: a flat disk and one with a hole in it. These are
separated with a small gap. The shaft passes through the hole with a
running fit and the end of the shaft rests on the flat part. When
oiling these, you should supply just enough oil to fill the space
between the two disks.
> In this case, I think I will avoid taking apart the dial mechanism, so
> I can't see the spring.
>From your description of where you are now with it, I'd say don't knock
success. :-)
Ted
Garry
"Feel" ok. If so it would be good for me.
Garry
On 10 Dec 2000 22:12:33 -0800, da...@cs.ubc.ca (Dave Martindale)
wrote:
Dave Martindale
and trying to figure out how to get the preload on the dial spring right
when putting it back together. I mentioned that I'd ended up giving it
3 turns and that it seemed to work well there.
Ted Edwards suggested that he would have used 4 turns, not 3, based
on experiences with other indicators. Recently, I went back to where
I bought the indicator and asked to borrow another one of the same
model. Then I measured the contact force on the tip throughout the
measuring range, using a postal scale. It ranged from 13-25 g. My own
reassembled indicator was 10-20 g, suggesting that mine was a bit
below normal.
So, this weekend I opened it up again. *This* time I clamped the
indicator's toolpost mounting bar in a vise and held the indicator
with its own mounting hardware, preventing me from dropping it.
I also used a micro Sharpie to mark how the gears meshed *before*
taking anything apart, so I could get it back to where it was at least.
Then I added about one more turn of preload, and put everything back
together. It isn't possible to add exactly one turn without disengaging
the crown gear from the dial pinion gear, which I didn't want to
attempt. So I just changed the engagement of the sector gear with
the crown gear by 5 teeth, which is one turn of the dial and a bit.
The contact force is now about 12-23 g, as close as I'm going to get
to the factory-fresh one. (One more turn would be too much). So
it appears Ted was right. Just playing with it, I can't see any
real difference in the behaviour between 3 and 4 turns.
For the curious: all the gears seem to have about 0.5 mm pitch.
The small end of the crown gear, which meshes with the sector gear,
has exactly 31 teeth (they're all visible for counting). The large
outer gear of the crown gear seems to have somewhere between 78 and
80 teeth. And the pinion gear on the dial shaft seems to have 12
teeth.
Dave